Author |
Topic: 2011 Postal Mixer (Read 16565 times) |
|
Hippo
Forum Guru
Arimaa player #4450
Gender: 
Posts: 883
|
 |
Re: 2011 Postal Mixer
« Reply #150 on: Oct 14th, 2011, 11:52am » |
 Quote  Modify
|
I want to announce the most nice game of the postal mixer. Of course you could not agree with me, but my game with Nombril was without big blunder (expect I will made one in next 6 turns). There were material combinations incluning goal threats in between, lasting above 10 halfturns.
The result was long time open with very interesting ending.
I think there was small inaccuracy just after the camel trade. I was much more afraid of e going to d3 early. h going to c2 let me save one Dog what was required to remain on the thin ice defending the goal attack around c3, but may be I am wrong.
Twice in the game (so far  ) there was sacrifice of second strongest piece just to gain time to solve situation in the opposing corner. And the capture was forced.
OK, if you have better candidates ... let you agitate for them
|
| « Last Edit: Oct 14th, 2011, 11:53am by Hippo » |
 IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
Re: 2011 Postal Mixer
« Reply #151 on: Oct 14th, 2011, 6:57pm » |
Quote Modify
|
on Oct 14th, 2011, 11:52am, Hippo wrote:| I want to announce the most nice game of the postal mixer. |
|
Oh, is the game over now?
[edit]
Ah, I see you announced goal in seven in the chat room. That is an impressive call to end an extremely dramatic game. If you are right, I will second your nomination for game of the tournament. I certainly couldn't tell what was going on, and I checked in frequently.
|
| « Last Edit: Oct 14th, 2011, 10:50pm by Fritzlein » |
IP Logged |
|
|
|
Hippo
Forum Guru
Arimaa player #4450
Gender:
Posts: 883
|
|
Re: 2011 Postal Mixer
« Reply #152 on: Oct 15th, 2011, 3:20pm » |
Quote Modify
|
Hmmm, my originally planned combination does not work ... it needed one defense step in the middle, what means enough delay to not work at all (another counterthreat).
So I have to find something else ... let us see result and resulting timing.
... may be it makes the game even more interesting
[edit]Now at 43s I am almost sure I will finish it in the called number of turns (47). But arimaa is full of surprises.[/edit]
[edit]At 43g I tought I could win till 47. At 43s I found a way the game could be delayed till 50 by counterthreats ... . Let us see what would be final result;)[/edit]
[edit]I have to stop making predictions.[/edit]
[edit]
So the game was rather OK till I have start talking about it rather than playing ... 42g Db1n Ce1n Ce2e Rd1n would be hard to response. I have given Eric too much free steps just to save my thinking time  . So consider the game the most beautifull just till 42g  .
[/edit]
|
| « Last Edit: Oct 20th, 2011, 4:59pm by Hippo » |
IP Logged |
|
|
|
Hippo
Forum Guru
Arimaa player #4450
Gender:
Posts: 883
|
|
Re: 2011 Postal Mixer
« Reply #153 on: Nov 17th, 2011, 11:40am » |
Quote Modify
|
It's finally over.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
Re: 2011 Postal Mixer
« Reply #154 on: Nov 17th, 2011, 12:20pm » |
Quote Modify
|
http://arimaa.com/arimaa/gameroom/comments.cgi?gid=202345
What a game to end with! The fact that the announced goal-in-seven turned into a goal-in-seventeen merely heightened the drama. There was a forum thread for "games that ought to be commentated"; this is my submission for that category. In particular, I wonder what bots would think of the endgame from move 40 onward. Is this an area in which modern software can surpass high-level human play, or is the situation still strategic enough that short-term tactics are overwhelmed by long-range human planning?
|
| « Last Edit: Nov 17th, 2011, 12:22pm by Fritzlein » |
IP Logged |
|
|
|
Hippo
Forum Guru
Arimaa player #4450
Gender:
Posts: 883
|
|
Re: 2011 Postal Mixer
« Reply #155 on: Nov 17th, 2011, 12:54pm » |
Quote Modify
|
Evaluation of my personal appearence on postal mixer:
I played 10 games, 7 of them were against opponents I am afraid of:
chessandgo, Adanac, Tuks, Fritzlein, rabbits, Nombril and unknown Donald.
I didn't liked my position with Donald when he timed out.
Daniel (Tuks) uses postal speed just for playing asynchroneous with opponent.
He does not use the extra time for thinking and thinks 5 minutes per
turn at most. He is that good player that even with these habbits he
is realy dangerous.
In the game we made equal trades of one horse pair and one dog pair.
Than I had to sacrifice rabbit to get camel freeze and not allow mine.
I got only dog for cat exchange from that and I had to release the camel.
I have made 3 steps for one step exchange to made stronger goal
protection and incidentaly that lead to Daniels overconfidence
to start attacking. That opened his back rows and increased
my goal chances what forced his phant to defend.
That allowed me to capture his cat. Daniel didn't notice he cannot
capture my rabbit in response due to horse two turn capture threat.
So finally I had horse and dog for three rabbits materially in the midgame.
I lost a concentration, what lead to loss of fourth rabbit.
Fortunately I had strong goal attack fixing most of opponent's pieces
for defense. Daniel missed that played a move and before I have noticed
it he finds the 2 turn loss and resigns.
Meanwhile I played an interesting combination with Gregory (rabbits), where my
rabbit was on my camel capture delaying duty allowing me to capture two horses
meanwhile. Gregory had to sacrifice his own rabbit to delay horse capture for free,
but I in thrill from my victory with Daniel I made a move with absolutely no concept.
I lost full turn for removing the rabbit from its duty.
I were not able to concentrate on the downhill battle after camel for horse exchange.
... I still hoped I could finally capture the other horse what lead to losing control
of my home traps and fast end.
Meanwhile Karl (Fritzlein) gained strong positional advantage in our game.
My cat was lost without interesting compensation.
I hoped for some disruption on the other wing, but it lead to loss control of one of my home traps.
When I lost control of the other home trap I have resigned.
I have lost without seeing any particular blunder. Fritzlein won this year postal mixer deservedly.
The game with Jean (chessandgo) was much more interesting for me.
He is that confident that he started cat elephant attack.
It is very unusual and I don't think it was correct.
I do think he would beat gold in silvers position, but I am not that strong as him.
I do think my position was still strong till 20s.
I had two options how to frame the rabbit, but I was that tired after long thinking in Eric's game
(camel exchange combination) that I played the 20s without deep checking the move is OK.
I have lost dog and whole position in several following moves.
The game with Greg (Adanac) we were several moves on treshold between playin home game and away game.
Finally I have decided to go for away game at 12s. I have strong grip on both Greg home traps while
he attacked my trap with horse. I have start feeling comfortable after taking horse hostage by camel at 20s.
I never considered my dog beeing hostaged by camel to be problem till I have senn the other dog
hostaged by horse during planning around 25g. The horse flip to f3 made me felt comfortable again.
28g and 29g made me a bit scared, but my calculations shown my goal attack is fast enough.
I don't know if I was underestimated in this game or Greg had just sacrified less time for thinking.
So the last game of postal mixer.
Eric (Nombril) declared himself to be away player letting me to frame his horse. I was waiting till his phant
is sufficiently tight to f3 to start M attacking his horse. It was in exact time.
Being one step slower Eric would finish blockade of my phant, having his phant free.
This way I made horse capture thread forcing Eric to let my phant breath.
This actually lead to sacrifice of my camel to get time to slove f3 trap problems entirely.
There was goal threat meanwhile gaining more time to got mc for MR, but when camels were traded
Eric had strong grip on the c3 trap. I was scared of ed3 move, but Eric went with horse to c2 before it.
This allowed me to save one dog for following goal defense. After the exchange was over, I had HC for
dr but with goal weak one wing with one horse totally ut of play (together with the other horse).
That of course meant Eric had one weak wing as well, but I was far behind with attacking there.
I had no chance to make a slow attack there and I had to sacrifice my second strongest piece for second time
forcing Eric's phant to cross wings and gain some tempo.
Another goal threat meanwhile gained enough time to capture dog for sacrified horse and to solve the
situation on the weaker wing. This forced dog for cat exchange and Eric's horse went to my first row.
The situation was very tense, I have seen goal attacks on both sides, but each attack lead to tragical
weakening of attackers goal defense. The rabbits around c3 were obstackles agianst my a file attack as well
as goal counterthreats. So I made capture threat with one stap defendable goal attack.
Eric started goal attack requiring full 4 steps for defense at 38s.
I used his horse as a brick in my defense.
At 40 we exchanged advanced rabbits and I started winning combination.
Eric had just cat, phant and 3 rank6 rabbits for defense.
There was goal in 7 at 41g, but there was defense with countergoal.
I have seen the countergoal at 42g so I made everything to prevent it. I was rather tired from the
long postal mixer season and I played quickly. Precise 42g would not end with rabbit on f1.
Cat on f2 would be appropriate. What I have missed is that silver could create another counterthreat
by pulling my dog from b2 to b3. Having my cat one step further, I could capture the g6 rabbit and made a
goal in one threat. This way I hade to sacrify two more steps to defend ra2.
My Horse was still safe as elephant with cat had not enough time to defend the goal attacks in case of capture.
I have tought about cat defense but was not sure what to do with the 4th step.
Rabbit to trap step surprised me. It prevented planned push to g7.
I still hope I had strong attack at 45g, but as I have seen 3 turn goal nice combination, I didn't payed
much attention to other moves. 10 minutes after playing the move, I have discovered that rabit
flip from g6 to h7 in front of cat sitting on h8 would not work.
I have seen re6 cg6 rg5 response and I understand that I have lost the game.
I think Eris'c positioning rabbit to f5 was weaker as it allowed me one more step after standing with cat at e6.
Morever I would be forced to leave rabbit on 6th rank were was the defense line of Eric.
I have decided to change the play speed.
I was happy Eric captured h6 rabbit rather than capturing the c4 Horse.
That allowed me to force him to capture the Horse in four steps ending et c5 rather to b4.
After the captures I were losing horse for dog materially, but with frozen horse as brick in my defense.
The only problem was opposition of cats. I could not allow Eric's cat to go to f4 so capture ending on f4 was must.
49s was risky, but without goal threats requiring 2 defense steps I would lost my dog followed by goal.
Fortunately I had one in hand.
50s surprised me as well, but I were in position to find nonlosing moves and there was no need to
find opponent's responses. So following two moves were forced.
Surprisingly a6 was better place for my rabbit than b7 as it does not allow capture to c7 defending d file.
I expected 53s Rc5 cd6 leading probably to difficult ellimination fight.
Unfortunately for arimaa, fortunately for me Eric haven't seen possible froze preventing Rc5 capture and that was the game over.
Uff, I have learned a lot.
|
|
IP Logged |
|
|
|
Hippo
Forum Guru
Arimaa player #4450
Gender:
Posts: 883
|
|
Re: 2011 Postal Mixer
« Reply #156 on: Nov 17th, 2011, 12:57pm » |
Quote Modify
|
on Nov 17th, 2011, 12:20pm, Fritzlein wrote:http://arimaa.com/arimaa/gameroom/comments.cgi?gid=202345
What a game to end with! The fact that the announced goal-in-seven turned into a goal-in-seventeen merely heightened the drama. There was a forum thread for "games that ought to be commentated"; this is my submission for that category. In particular, I wonder what bots would think of the endgame from move 40 onward. Is this an area in which modern software can surpass high-level human play, or is the situation still strategic enough that short-term tactics are overwhelmed by long-range human planning? |
|
There was probably goal in 7 or 8, but I have missed it.
The goal in 17 was actually Eric's fault not my good play.
And Yes, I made a lot of positions to test my bot on. I am happy this is permited now .
|
| « Last Edit: Nov 17th, 2011, 1:08pm by Hippo » |
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
Re: 2011 Postal Mixer
« Reply #157 on: Nov 17th, 2011, 1:22pm » |
Quote Modify
|
on Nov 17th, 2011, 12:57pm, Hippo wrote:| There was probably goal in 7 or 8, but I have missed it. The goal in 17 was actually Eric's fault not my good play. |
|
Do you mean that your bot can verify that you had a won position, then a lost position, and then a won position again? What is the longest goal it can announce from a position in this game?
|
|
IP Logged |
|
|
|
Hippo
Forum Guru
Arimaa player #4450
Gender:
Posts: 883
|
|
Re: 2011 Postal Mixer
« Reply #158 on: Nov 17th, 2011, 1:40pm » |
Quote Modify
|
on Nov 17th, 2011, 1:22pm, Fritzlein wrote:
Do you mean that your bot can verify that you had a won position, then a lost position, and then a won position again? What is the longest goal it can announce from a position in this game? |
|
Surprisingly I have not tested it yet .
I don't think so, but I suppose I will use this game for testing during it's development during 2012 season.
There was a lot of goal threats from around 38 so the search tree could be cut a lot. I would be interested what briareus/quad/marwin thinks after 4 hours of thinking at each move in the given range.
|
|
IP Logged |
|
|
|
rbarreira
Forum Guru
Arimaa player #1621
Gender:
Posts: 605
|
|
Re: 2011 Postal Mixer
« Reply #159 on: Nov 17th, 2011, 6:36pm » |
Quote Modify
|
I ran briareus on some positions from that game, disabling unsafe pruning stuff such as null-move pruning in order to have more sure analysis (but of course some bugs might still be present).
With the disclaimer out of the way, for now I can say all the silver positions from 54s to 57s are indeed lost (so every gold move from 54 on was winning, even if not the fastest win).
46s rf5s rf4w re4s re3w was a goal in 2
52s eb4n eb5n hh1w rh5s might be worth looking at.
I'll post more analysis later if I have time...
|
|
IP Logged |
|
|
|
rbarreira
Forum Guru
Arimaa player #1621
Gender:
Posts: 605
|
|
Re: 2011 Postal Mixer
« Reply #160 on: Nov 17th, 2011, 6:43pm » |
Quote Modify
|
46g Cg7w Cf7w Ce7s Ce6s avoids the goal (or postpones it at least).
|
|
IP Logged |
|
|
|
rbarreira
Forum Guru
Arimaa player #1621
Gender:
Posts: 605
|
|
Re: 2011 Postal Mixer
« Reply #161 on: Nov 17th, 2011, 6:48pm » |
Quote Modify
|
on Nov 17th, 2011, 1:40pm, Hippo wrote:
I would be interested what briareus/quad/marwin thinks after 4 hours of thinking at each move in the given range. |
|
I may run some long analyses on a few particular moves that might be interesting (what would be good candidates?). To run 4-hour analyses on many moves I'd probably need to get paid to have incentive to do it  (and to pay the electricity bills).
|
| « Last Edit: Nov 17th, 2011, 6:55pm by rbarreira » |
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
Re: 2011 Postal Mixer
« Reply #162 on: Nov 18th, 2011, 7:37am » |
Quote Modify
|
on Nov 17th, 2011, 6:48pm, rbarreira wrote:| I may run some long analyses on a few particular moves that might be interesting (what would be good candidates?). To run 4-hour analyses on many moves I'd probably need to get paid to have incentive to do it (and to pay the electricity bills). |
|
Speaking of financial incentives, how much do you charge for a copy of briareus for my own use in home analysis? I have Bomb2005, but it is rather buggy in the endgame; it seems to get hung up in repetition and to have some issues with positions where passing steps is correct. I'm ready to take a step up with a bot that analyzes endgames better, especially since that is a weak part of my game.
|
|
IP Logged |
|
|
|
rbarreira
Forum Guru
Arimaa player #1621
Gender:
Posts: 605
|
|
Re: 2011 Postal Mixer
« Reply #163 on: Nov 18th, 2011, 9:32am » |
Quote Modify
|
on Nov 18th, 2011, 7:37am, Fritzlein wrote:
Speaking of financial incentives, how much do you charge for a copy of briareus for my own use in home analysis? I have Bomb2005, but it is rather buggy in the endgame; it seems to get hung up in repetition and to have some issues with positions where passing steps is correct. I'm ready to take a step up with a bot that analyzes endgames better, especially since that is a weak part of my game. |
|
I never thought about that... The bot is in a rather crude state right now (with the eval getting completely replaced) which makes me wary to sell it, but I suppose it could still be useful to some people.
On the other hand I'm not sure I want to get into the whole mess of selling something in light of the restrictive Arimaa license. But to tell the truth I'm not sure that the patent on Arimaa has any restriction on me as I do not live in the U.S. and AFAIK Arimaa is not patented in other countries (and selling an Arimaa bot certainly does not violate copyright). Any comments Omar?
I'll PM you later Fritz.
|
| « Last Edit: Nov 18th, 2011, 9:33am by rbarreira » |
IP Logged |
|
|
|
Hippo
Forum Guru
Arimaa player #4450
Gender:
Posts: 883
|
|
Re: 2011 Postal Mixer
« Reply #164 on: Nov 18th, 2011, 7:21pm » |
Quote Modify
|
I would like to know analysis of position prior 42g.
The Cat invasion to g6 when elephant had to solve west problems seems to me is forced win. Unfortunately I have lost at least one deciding step on 42g forcing me to lose 2 more steps next move ... .
|
| « Last Edit: Nov 18th, 2011, 7:23pm by Hippo » |
IP Logged |
|
|
|
|
Home
Help
Search
Members
Login
Register
Reply
Notify of replies
Send Topic
Print